Humanity Says Yes
Our modification of environments for machine compatibility raises critical concerns of our ability to take decisions.
Consider teaching lawn mowing to a robot. A long wire laid around the lawn’s borders, and the mower is let loose. The machine will criss-cross the area within the borders until the task is done. Congratulations, you have now adapted your lawn into an object that the machine can understand.
I recently read a few snippets from Luciano Floridi’s book, The Fourth Revolution. An interesting quote stood out to me: “We have been enveloping the world around information technologies for decades without fully realising it.”
The idea seems pertinent in an era where every action is heavily data-driven. We are essentially enveloping every action in a dense, encrypted and secure data layer. We are at a stage where deciphering, retrieving and modifying our own data is impossible without the intervention of memory, speed and pattern recognition.
The best AI systems currently are more inflexible than the most disagreeable human being. This is fostering a practice of designing, performing and running iterations in a way that a computer system can understand. I have often heard the phrase ‘the computer can’t do it’ — which essentially translates to “I don’t know how to do it on the computer.”
The best AI — unless by deux ex machina achieves sentience — will always obey its code.
This way, the minority whose rules apply are those designing machines in the first place.
A robot assuming a human’s responsibility seems far away. Most AI applications are not similar to human beings either in their functions or their faculties. It is for this reason that we are transforming the very fabric of our lives into a format understood by machines.
If we are to effectively debate adapting our world into a one mediated by machines, this has to occur at the design phase. It is important to have our assets and refusals before designing newer systems so that there is room for flexibility.
Perhaps this will require a convergence of different fields as productivity and profit spheres are currently ruling the design requirements of artificial intelligence. I anticipate a system which can take into account aspects of justice, freedom and compassion, to name a few.
Before the phrase ‘the computer can’t do it’ becomes mainstream, it is time humanity says yes to decide on a framework which can make AI adaptable to our world instead of the other way around.
